It is estimated that more than half of all eukaryotic proteins are glycoproteins, which implies that specific amino acid side chains are chemically modified with carbohydrates. The most abundant form of these modifications is asparagine-linked (“N-linked”) glycosylation, which affects a multitude of cellular functions that range from protein folding, quality control, sorting and secretion to organism development and host-pathogen interactions. Asparagines facing the lumen of the endoplasmic reticulum (ER) are specifically glycosylated when located in the consensus “sequon” Asn-X-Ser/Thr, where X may be any amino acid except proline. The reaction takes place at the membrane surrounding the ER and is catalysed by the enzyme oligosaccharyltransferase (OST), a hetero-oligomeric protein complex embedded in the ER membrane of higher eukaryotes (see FIG. 1b). A hallmark of N-linked glycosylation is its broad specificity with respect to the polypeptide substrate, which is a direct consequence of the short recognition sequon. This characteristic distinguishes OST from glycosyltransferases that modify serine or threonine residues (O-linked glycosylation) and exhibit a higher specificity for their protein substrates.
The key step in OST-catalysed glycosylation is the formation of an N-glycosidic linkage between the amide nitrogen of an acceptor asparagine and the C1 carbon of the first saccharide moiety of a lipid-linked oligosaccharide (LLO) donor (see FIG. 1a). This results in the en bloc transfer of the oligosaccharide onto the acceptor asparagine. Details of the underlying reaction mechanism are poorly understood. This is due to the absence of structural insight into OST at high resolution, but also to the complex chemical nature of the LLO substrate, its low abundance in biological samples, and its insolubility in water. In contrast, crystal structures of various soluble O-glycosyltransferases have been published and their reaction mechanisms were investigated in great detail. For OST, the currently accepted model suggests that glycosylation sequons are recognized when located in unfolded protein segments, which can occur during protein translocation into the ER or after translocation is completed. The central catalytically active component within OST is the STT3 subunit, whereas the other subunits are thought to assist and refine the process by facilitating OST complex assembly or by interacting with a subset of acceptor proteins or the LLO substrate, leading to an increased number of accessible and modified glycosylation sites.
N-linked glycosylation is not restricted to eukaryotes. Homologous processes are found in archaea and in defined taxa of proteobacteria. However, prokaryotes and eukaryotic kinetoplastids contain a single-subunit OST enzyme that is homologous to the STT3 subunit of higher eukaryotes. The best-studied prokaryotic N-glycosylation process is mediated by the protein glycosylation locus pgl from the bacterium Campylobacter jejuni (Szymanski et al. (1999) Molecular Microbiology 32, 1022-1030). The locus contains an integral membrane protein termed PglB that shares significant sequence similarity with eukaryotic STT3, suggesting a common membrane topology and reaction mechanism (see FIG. 1b). This gene cluster is sufficient for catalyzing protein glycosylation when transferred into Escherichia coli cells. OST-catalysed prokaryotic protein glycosylation of sequon-containing protein substrates is an economic, effective and convenient way of glycosylating recombinantly produced proteins (Wacker et al. (2002), Science 298, 1790-1793). N-linked protein glycosylation can be engineered with diverse O antigen lipopolysaccharide structures of non-C. jejuni origin in E. coli (Feldman et al. (2005) PNAS 102(8), 3016-3021), thus allowing for the prokaryotic transfer of eukaryotic N-glycans to recombinant protein substrates. Glover et al. (2005, Chemistry & Biology 12, 1311-1315) demonstrated for the first time the in vitro protein glycosylation using E. coli cell membranes comprising overexpressed PglB and an undecaprenyl pyrophosphate bound oligosaccharide. In 2006 Kowarik et al. (2006, EMBO J. 25(9), 1957-1966) further defined the bacterial N-glycosylation site consensus sequence by showing that the substrate specificity of bacterial OST is extended to a negatively charged amino acid in the −2 position of the acceptor asparagine, resulting in the consensus sequon Asp/Glu-X1-Asn-X2-Ser/Thr (wherein X1 and X2 are both not proline; SEQ ID NO: 3). By using a peptide substrate library, Chen et al. (2007, Biochemistry 46, 5579 -5585) confirmed the necessity for a negative charge in the −2 position of the acceptor asparagine and identified the sequence DQNAT (SEQ ID NO: 4) as the optimal substrate for C. jejuni PglB.
From the above it follows that the prokaryotic oligosaccharyltransferase (OST) has a broad, because small sequon-based specificity for protein substrates and can be used to transfer eukaryotic, prokaryotic as well as synthetic N-glycans. Essentially the prokaryotic OST-based N-glycosylation system requires three components, (a) an oligosaccharide donor, preferably a lipid- or undecaprenyl pyrophosphate-linked oligosaccharide donor, (b) a prokaryotic oligosaccharyltransferase (OST), (c) a potential consensus sequence motif polypeptide substrate and last but not least a suitable physiological microenvironment, e.g. cell membranes in vitro or in vivo.
The problem underlying the present information is that one cannot predict or design components essential for the very versatile prokaryotic OST-based N-glycosylation system beyond the information given for already known OST-components. In addition, there is no clue what structural requirements a potential OST glycosylation inhibitor must have. Such inhibitors would be expected to have pronounced biological effects and could be of great medical, diagnostic and scientific value. The other problem is that up to now it had not been possible to provide a three-dimensional model of the catalytic domain and the polypeptide binding site of an OST that could have provided the scientific community with insight regarding the possible variation of the components involved in OST-mediated glycosylation.
The above problems have been solved by the provision of the three-dimensional X-ray structure of a bacterial OST, the PglB protein from Campylobacter lari (SEQ ID NO:1; sharing 56% sequence identity with PglB of the C. Jejuni) in complex with the acceptor hexapeptide DQNATF (SEQ ID NO: 5). C. lari PglB (SEQ ID NO:1) is active when co-expressed with the C. jejuni pgl cluster in E. coli cells, as evidenced by glycosylation of an acceptor protein containing a consensus sequon (see FIG. 2). For its structural analysis, C. lari PglB (SEQ ID NO:1) was co-crystallized with the hexapeptide DQNATF (SEQ ID NO: 5) that contains the sequon glycosylated in the in vivo assay, which had been identified as the optimal acceptor sequence for C. jejuni PglB (see Chen et al. above). The structure of C. lari PglB (712 amino acid residues, SEQ ID NO: 1) was determined using a combination of experimental phasing and molecular replacement, making use of the previously determined structure of the periplasmic domain of C. jejuni PglB. Co-crystals of PglB (SEQ ID NO:1) were small, fragile, and diffracted X-rays anisotropically, with the best native data extending to 3.4 Å resolution. The structure was refined to R/Rfree values of 23.8 and 27.1%, respectively (Table 2). Further details of the structure of C. lari PglB (SEQ ID NO:1) are provided in the experimental section below.
This new three-dimensional structure provides insight into the molecular basis for sequon recognition and reveals a catalytic site that is formed by the transmembrane domain of the protein and features conserved, acidic side chain residues and a bound divalent cation. These results suggest for the first time a mechanism for amide nitrogen activation and glycosylation and provide a reasoned approach for identifying and designing new oligosaccharide donors, new oligosaccharyltransferase variants (OST), new consensus sequence motif polypeptides, as well as OST glycosylation inhibitors, all of which have utility in recombinant glycoprotein production, diagnostics, medicine and as scientific tools.
In view of the above, a first aspect of the invention relates to a method for identifying a potential component for the oligosaccharyltransferase (OST)-catalysed asparagine-linked (“N-linked”) glycosylation selected from the group consisting of                (a) a potential oligosaccharide donor, preferably a lipid-linked oligosaccharide (LLO) or an undecaprenyl pyrophosphate bound oligosaccharide donor,        (b) a potential oligosaccharyltransferase (OST),        (c) a potential consensus sequence motif polypeptide, and        (d) a potential glycosylation inhibitor, comprising the steps of        (i) using the atomic coordinates of Table 1, preferably +2, more preferably +1.5, most preferably +1.0 Å root mean square deviation (rmsd) from the backbone atoms, for generating a three-dimensional model of the catalytic domain of the oligosaccharyl-transferase (OST) of Campylobacter lari (SEQ ID NO:1), comprising at least one, two, three, four, five, six, seven, most preferably all amino acids D56, R147, D154, D156, E319, R375, Y468, and H485 of SEQ ID NO:1, and/or, preferably and        (ii) using the atomic coordinates of Table 1, preferably +2, more preferably +1.5, most preferably +1.0 Å root mean square deviation (rmsd) from the backbone atoms, for generating a three-dimensional model of the polypeptide binding site of the oligosaccharyltransferase (OST) of Campylobacter lari (SEQ ID NO:1), comprising at least one, two, three, four, five, most preferably all of amino acids M318, R331, W463, W464, D465, and I572 of SEQ ID NO:1,        (iii) preferably performing whole body translations and/or rotations on the coordinates of the amino acids of the three-dimensional models of (i) and/or (ii),        (iv) using said three-dimensional model(s) of (i), (ii) and/or (iii) for designing or selecting at least one of potential components (a) to (d),        (v) providing at least one of said potential components (a) to (d), and        (vi) contacting at least one of said potential components (a) to (d) with the further functional components necessary for an oligosaccharyltransferase (OST)-catalysed asparagine-linked (“N-linked”) glycosylation,        (vii) identifying a functional component selected from the group consisting of                    (A) a functional oligosaccharide donor, preferably a functional lipid-linked oligosaccharide (LLO) donor or an undecaprenyl pyrophosphate bound oligosaccharide donor,            (B) a functional oligosaccharyltransferase (OST),            (C) a functional consensus sequence motif polypeptide, and            (D) a functional glycosylation inhibitor.                        
In a preferred embodiment, in step (ii) the three-dimensional model of the polypeptide binding site of the oligosaccharyltransferase (OST) of Campylobacter lari (SEQ ID NO:1) comprises at least two, preferably at least three, more preferably at least four, most preferably all of amino acids M318, A331, W463, W464, D465 and I572 of SEQ ID NO: 1.
The atomic coordinates of Table 1 for use in the methods of the present invention are shown in FIG. 7. The X-ray coordinates of the OST of C. lari (SEQ ID NO:1), in particular of the catalytic site and/or the polypeptide binding site complexed with the optimized polypeptide substrate DQNAT (SEQ ID NO: 4) provide the skilled person with the three-dimensional information necessary for identifying a potential component for OST catalysis and catalytic inhibition. The spatial restrictions in combination with the chemical functional nature of the individual atoms of the amino acids involved in catalytic action and polypeptide binding, e.g. electron densities, position of van der Waals forces, ionic interactions, hydrophobic interactions, etc., inform the person skilled in computer-assisted molecular modelling of the structural and spatial prerequisites of (a) a potential oligosaccharide donor, preferably a lipid-linked oligosaccharide (LLO) or an undecaprenyl pyrophosphate bound oligosaccharide donor; (b) a potential oligosaccharyltransferase (OST), (c) a potential consensus sequence motif polypeptide, and/or (d) a potential glycosylation inhibitor.
As described before bacterial OSTs have a broad specificity for oligosaccharide donor molecules. With the coordinate information and the method of the invention, the repertoire of useful oligosaccharide donors can be rationally designed and extended without having to revert to trial and error synthetic strategies. Also, the OST itself can be rationally varied without rendering the catalytic site and the polypeptide binding pocket non-functional. This OST variation is useful, for example for modifying the catalytic potential, polypeptide substrate specificity and/or oligosaccharide donor specificity of OSTs. Moreover, the consensus motif of the polypeptide oligosaccharide acceptor can be rationally varied and designed, thus leading to the broadening of the OST utility such as e.g. glycosylation of eukaryotic sites. Last but not least, the inventive three-dimensional X-ray model provides an excellent basis for designing potential glycosylation inhibitors, which can be expected to be physiologically active by interrupting, modifying or slowing OST activity. These inhibitors have a great potential for providing scientific, diagnostic and therapeutic tools.
The term “root mean square deviation” or “rms deviation” or “rmsd” means the square root of the arithmetic mean of the square of the deviations from the mean. In the context of atomic objects the numbers are given in angstroms (Å). It is a way of expressing the deviation or variation from a trend or object.
The method of the invention comprises the step of using the atomic coordinates of Table 1, preferably +2, more preferably +1.5, most preferably +1.0 Å root mean square deviation (rmsd) from the backbone atoms, for generating a three-dimensional model of the catalytic domain of the oligosaccharyltransferase (OST) of Campylobacter lari (SEQ ID NO:1), comprising at least one, two or three, preferably at least four, more preferably at least five or six, most preferably seven or all of amino acids D56, R147, D154, D156, E319, R375, Y468, and H485 of SEQ ID NO:1, and/or, preferably and using the atomic coordinates of Table 1, preferably +2, more preferably +1.5, most preferably +1.0 Å root mean square deviation (rmsd) from the backbone atoms, for generating a three-dimensional model of the polypeptide binding site of the oligosac-charyltransferase (OST) of Campylobacter lari (SEQ ID NO:1), comprising at least one or two, preferably at least three, more preferably at least four or five, most preferably all of amino acids M318, R331 (or A331), W463, W464, D465, and I572 of SEQ ID NO:1.
At the polypeptide binding site W463, W464 and D465 of SEQ ID NO:1 form strong hydrogen bonds to the β-hydroxyl group of T in the bound acceptor sequon. These residues strongly contribute to acceptor sequon binding and are the reason why an S or T is located at the +2 position of the acceptor asparagine (N—X—S/T). R331 of SEQ ID NO:1 forms a salt bridge to the negatively charged D of the acceptor polypeptide and therefore contributes to acceptor sequon binding. R331 of SEQ ID NO:1 is responsible for the requirement of a negatively charged amino acid in the −2 position of the acceptor asparagine and is responsible for the extension of the consensus sequon for bacterial N-linked protein glycosylation (D/E-X1—N—X2—S/T; SEQ ID NO: 3) (FIG. 4a). R331 of SEQ ID NO:1 is only conserved in bacterial OSTs and can serve as a target for changing the substrate specificity of PglB towards the recognition of eukaryotic glycosylation sites. In fact, when R331 of SEQ ID NO:1 is mutated to A, the resulting glycosylation site AQNAT (SEQ ID NO: 8) in an acceptor protein, e.g. modified scFv fragment 3D5 originally containing the sequon DQNAT (SEQ ID NO: 4), can be glycosylated, whereas this site does not serve as a substrate for the wild type enzyme (FIG. 4b).
At the catalytic site, D56, D154 and E319 of SEQ ID NO:1 seem responsible for the coordination of the bound divalent metal ion, which is essential for catalysis. More importantly, D56 and E319 of SEQ ID NO:1 form hydrogen bonds to the amido group of the asparagine side chain of the bound acceptor sequon. This interaction causes a rotation of the C—N bond in the amido group, which is important for the nucleophilic activation of the nitrogen. D156 and R147 of SEQ ID NO:1 stabilize the hydrogen bonding network and R375 of SEQ ID NO:1 complexes the negative charge of one of the phosphates of a bound lipid-linked oligosaccharide (LLO).
For example, in the polypeptide binding site W463, W464 and D456 of SEQ ID NO:1 could be replaced by two H and an E (WWD→HHE). To keep the requirement of the negative charge in the −2 position of the acceptor asparagine, R331 of SEQ ID NO:1 could be replaced by K. To create the requirement of a positive charge in this position, R331 of SEQ ID NO:1 could be replaced by D or E. To overcome the requirement of the negative charge in the −2 position of the acceptor asparagine, R331 of SEQ ID NO:1 can be replaced by A (FIG. 4). At the catalytic site D154 and D156 of SEQ ID NO:1 can be replaced by E each. E319 of SEQ ID NO:1 could be replaced by D and D56 of SEQ ID NO:1 could be replaced by E, respectively (see below). To modify PglB activity, for example, D56, D154 and E319 of SEQ ID NO:1 could be replaced by alanines or the corresponding amino function (N or Q). H485 of SEQ ID NO:1 could be replaced by W as W appears in eukaryotic OSTs at this position.
The structure coordinates of PglB from C. lari (SEQ ID NO:1) as listed in Table1 in FIG. 7 are accessible for download to the skilled person from the pdb database. (The research collaboratory for structural bioinformatics (RCSB) Protein Database (PDB)). With the help of a computer and freely accessible structure programs such as PyMOL or commercially available structure programs the skilled person can easily generate three-dimensional models useful for the claimed method.
Those of skill in the art will understand that a set of structure coordinates for a protein, protein/substrate or protein/inhibitor complex or a portion thereof is a relative set of points that defines a shape in three dimensions. Thus, it is possible that an entirely different set of coordinates could define a similar or identical shape. For this reason it is preferred to perform whole body translations and/or rotations on the coordinates of the amino acids of the three-dimensional models (i) and/or (ii) obtained from the atomic coordinates of Table 1. These variations in coordinates may be generated by mathematical manipulations of the structure coordinates, for example manipulation by crystallographic permutations of the structure coordinates, fractionalization or matrix operations to sets of structure coordinates or any combination of the above.
Next, the above generated three-dimensional models are used for designing or selecting at least one of the potential components of the OST glycosylation or a potential inhibitor. Various computational analyses are necessary to determine whether a molecule such as a specific oligosaccharide donor, a modified OST, an oligosaccharide acceptor consensus sequence motif polypeptide or a glycosylation inhibitor is sufficiently designed to rationally predict its functionality in the OST-catalysed reaction. Spatial, functional and chemical considerations such as the nature, position of atoms, rotational degree of freedom, electron density, steric hindrance, van der Waals, ionic and hydrophobic interactions, etc. will have to be made. Such analysis may be carried out conveniently on computers by standard software applications such as CCP4 (COLLABORATIVE COMPUTATIONAL PROJECT, NUMBER 4. 1994. “The CCP4 Suite: Programs for Protein Crystallography”. Acta Cryst. D50, 760-763.)
Current computational molecular similarity applications permit comparisons between different structures, different conformations of the same structure and different parts of the same structure. The comparison procedure is typically divided into four steps: (1) loading the structural information, (2) defining atom equivalence on these structures, (3) performing a fitting (superimposition) operation and (4) analysis of the results. Each structure is identified by a name. One structure is then identified as the potential OST component or OST inhibitor (i.e. the target or fixed structure), all remaining structures are working structures (i.e. moving structures). When a rigid fitting method is used the working structure is translated and rotated to obtain an optimum fit (spatial and functional complementarity) with the target structure, e.g. an OST inhibitor is the fixed structure and the OST amino acids are translated and rotated to obtain the optimum fit. The fitting operation uses an algorithm that computes the optimum translation and rotation to be applied to the moving structure, such that the rmsd of the fit over the specified pairs of equivalent atom is an absolute minimum. After superimposition of the two structures an rmsd value can be calculated for specific sets of equivalent atoms.
The potential functional component or inhibitor of the OST reaction selected or designed according to the inventive method as described above will provide the skilled person with a reasonable expectation of success when verifying its functionality in a routine OST activity assay. For said purpose the potential component must be provided by purchase, modification of purchased materials, chemical and/or recombinant synthesis, etc. This potential component will then have to be contacted with the further functional components necessary for an OST-catalysed asparagine-linked (“N-linked”) glycosylation, of course under conditions that allow for OST activity. Preferred OST activity assays are described in (1) 2005 Chemistry & Biology 12, 1311-1315, (2) 2006 Science 314, 1148-1150, (3) 2007 Biochemistry 46, 5579-5585, (4) 2007 Glycobiology 11, 1175-1182 and (5) 2011 Glycobiology 5, 575-583. Whether or not the potential components or inhibitor is OST active is preferably verified in comparison to positive or negative standards. For example, functionality of the OST assay is established with a known oligosaccharide acceptor polypeptide, e.g. the hexapeptide DQNATF (SEQ ID NO: 5), and then the potential functional consensus sequence motif polypeptide is substituted for the hexapeptide and glycosylation of the substitute polypeptide is determined. This simple OST assay system can be adapted for identifying any functional OST components, preferably one selected from the group consisting of (A) a functional oligosaccharide donor, preferably a functional lipid-linked oligosaccharide (LLO) or an undecaprenyl pyrophosphate bound oligosaccharide donor, (B) a functional oligosaccharyltransferase (OST), (C) a functional consensus sequence motif polypeptide, and (D) a functional glycosylation inhibitor.
In a second aspect, the present invention relates to a method for designing a potential component for the oligosaccharyltransferase (OST)-catalysed asparagine-linked (“N-linked”) glycosylation selected from the group consisting of (a) a potential oligosaccharide donor, preferably a lipid-linked oligosaccharide (LLO) or an undecaprenyl pyrophosphate bound oligosaccharide donor, (b) a potential oligosaccharyltransferase (OST), (c) a potential consensus sequence motif polypeptide, and (d) a potential glycosylation inhibitor, comprising the steps of
(i) using the atomic coordinates of Table 1, preferably +2, more preferably +1.5, most preferably +1.0 Å root mean square deviation (rmsd) from the backbone atoms, for generating a three-dimensional model of the catalytic domain of the oligosaccharyl-transferase (OST) of Campylobacter lari (SEQ ID NO:1), comprising at least one, two or three, preferably at least four or five, more preferably at least six or seven, most preferably all amino acids D56, R147, D154, D156, E319, R375, Y468, and H485 of SEQ ID NO:1, and/or, preferably and
(ii) using the atomic coordinates of Table 1, preferably +2, more preferably +1.5, most preferably +1.0 Å root mean square deviation (rmsd) from the backbone atoms, for generating a three-dimensional model of the polypeptide binding site of the oligo-saccharyltransferase (OST) of Campylobacter lari (SEQ ID NO:1), comprising at least one or two, preferably at least three, more preferably at least four or five, most preferably all of amino acids M318, R331, W463, W464, D465, and I572 of SEQ ID NO:1,
(iii) preferably performing whole body translations and/or rotations on the coordinates of the amino acids of the three-dimensional models of (i) and/or (ii),
(iii.1) using said three-dimensional model of (i), (ii) and/or (iii) for assessing the stereochemical complementarity between said three-dimensional model(s) (i), (ii) and/or (iii) and a known or potential component for the oligosaccharyltransferase (OST)-catalysed asparagine-linked (“N-linked”) glycosylation selected from an oligosaccharide donor, preferably a lipid-linked oligosaccharide (LLO) donor or an undecaprenyl pyrophosphate bound oligosaccharide donor, a consensus sequence motif polypeptide, and a potential glycosylation inhibitor, or
(iii.2)varying at least one amino acid in said three-dimensional model of (i), (ii) and/or (iii) and using said varied three-dimensional model of (i), (ii) and/or (iii) for assessing the stereochemical complementarity between said three-dimensional models (i), (ii) and/or (iii) and a known or potential component for the oligosaccharyltransferase (OST)-catalysed asparagine-linked (“N-linked”) glycosylation selected from an oligosaccharide (LLO) donor, preferably a lipid-linked oligosaccharide (LLO) donor or an undecaprenyl pyrophosphate bound oligosaccharide donor, a consensus sequence motif polypeptide, and a potential glycosylation inhibitor,
(iv) optimizing said sterochemical complementarity in an iterative approach by observing changes in the three-dimensional model of (iii.1), (iii.2) or the component for the oligosaccharyltransferase (OST)-catalysed asparagine-linked (“N-linked”) glycosylation,
(v) designing a potential component selected from (a) to (d) which optimizes said stereo-chemical complementarity of said three-dimensional model(s) and potential component,
(vi.1)optionally providing the optimized potential component, and
(vi.2)contacting at least one of said potential components (a) to (d) with the further functional components necessary for an oligosaccharyltransferase (OST)-catalysed asparagine-linked (“N-linked”) glycosylation,
(vi.3)identifying a functional component selected from the group consisting of (A) a functional oligosaccharide donor, preferably a functional lipid-linked oligosaccharide (LLO) or an undecaprenyl pyrophosphate bound oligosaccharide donor, (B) a functional oligosaccharyltransferase (OST), (C) a functional consensus sequence motif polyeptide, and (D) a functional glycosylation inhibitor.
In a preferred embodiment, in step (ii) the three-dimensional model of the polypeptide binding site of the oligosaccharyltransferase (OST) of Campylobacter lari (SEQ ID NO:1) comprises at least two, preferably at least three, more preferably at least four, most preferably all of amino acids M318, A331, W463, W464, D465 and I572 of SEQ ID NO:1.
This method is basically very similar to the method of the first aspect except that in the method directly above the potential OST component is designed by optimizing its stereochemical complementarity to the three-dimensional models with or without whole body translations and rotations in an iterative approach by observing changes in the three-dimensional models or the component for the oligosaccharyltransferase (OST)-catalysed asparagine-linked (“N-linked”) glycosylation, when varying at least one amino acid in at least one of said three-dimensional models.
Once the designed potential OST component is selected based on its optimized stereo-chemical complementarity to said three-dimensional model(s) it can optionally be verified in an OST assay, preferably by providing the optimized potential component (by chemical and/or recombinant synthesis, purchase, modification of known compounds, etc.), contacting said optimized potential component with the further functional components necessary for an oligosaccharyltransferase (OST)-catalysed asparagine-linked (“N-linked”) glycosylation. In a last optional step the functional component of the OST reaction or an inhibitor thereof is identified by its impact on the OST reaction. Typically, positive and negative reference components are used to verify OST assay activity.
In a preferred embodiment of the methods of the present invention for identifying or designing potential OST components, the specific three-dimensional catalytic site model of step (i) further comprises one or more, preferably at least 5, more preferably at least 10, most preferably all of the amino acids selected from the group having residues located within Van der Waals distance to the bound peptide of SEQ ID NO: 2, preferably selected from those within a distance of 5 Å to said peptide, more preferably selected from the group consisting of T53, T54, N55, D56, N146, R147, Y152, E315, T316, I317, M318, E319, V320, N321, R331, L374, R375, Y433, S435, V438, W463, W464, D465, G482, H485, I572, V575 of SEQ ID NO:1.
In a further aspect the present invention relates to a machine-readable medium comprising, e.g. storing                (i) the atomic coordinates of Table 1, preferably +2, more preferably +1.5, most preferably +1.0 Å root mean square deviation (rmsd) from the backbone atoms, preferably comprising at least one, two or three, preferably at least four or five, more preferably at least six or 7, most preferably all of amino acids D56, R147, D154, D156, E319, R375, Y468, and H485 of SEQ ID NO:1, and/or, preferably and        (ii) the atomic coordinates of Table 1, preferably +2, more preferably +1.5, most preferably +1.0 Å root mean square deviation (rmsd) from the backbone atoms, for generating a three-dimensional model of the polypeptide binding site of the oligosaccharyltransferase (OST) of Campylobacter lari (SEQ ID NO:1), comprising at least one or two, preferably at least three, more preferably at least four or five, most preferably all of amino acids M318, R331 (or A331) , W463, W464, D465, and I572 of SEQ ID NO:1,        (iii) preferably the atomic coordinates of (i) or (ii) modified by performing whole body translations and/or rotations on said coordinates.        
The above medium is particularly useful for a variety of purposes such as computer-assisted drug design, drug discovery and the X-ray crystallographic analysis of OSTs from other bacteria.
In the following the present invention will be further illustrated with reference to specific embodiments and experiments which are not intended to be interpreted as limiting the scope of the invention as presented by the appended claims.
SEQ ID NO: 1 lists the 712 amino acids of the oligosaccharyltransferase (OST) of C. lari PglB.
MELQQNFTDNNSIKYTCILILIAFAFSVLCRLYWVAWASEFYEFFFNDQL MITTNDGYAFAEGARFDMIAGFHQPNDLSYFGSSLSTLTYWLYSILPFSF ESIILYMSTFFASLIVVPIILIAREYKLTTYGFIAALLGSIANSYYNRTM SGYYDTDMLVLVLPMLILLTFIRLTINKDIFTLLLSPIFIMIYLWWYPSS YSLNFAMIGLFGLYTLVFHRKEKIFYLAIALMIIALSMLAWQYKLALIVL LFAIFAFKEEKINFYMIWALIFISISILHLSGGLDPVLYQLKFYVFKASD VQNLKDAAFMYFNVNETIMEVNTIDPEVFMQRISSSVLVFILSFIGFILL CKDHKSMLLALPMLALGFMALRAGLRFTIYAVPVMALGFGYFLYAFFNFL EKKQIKLSLRNKNILLILIAFFSISPALMHIYYYKSSTVFTSYEASILND LKNKAQREDYVVAWWDYGYPIRYYSDVKTLIDGGKHLGKDNFFSSFVLSK EQIPAANMARLSVEYTEKSFKENYPDVLKAMVKDYNQTSAKDFLESLNDK NFKFDTNKTRDVYIYMPYRMLRIMPVVAQFANTNPDNGEQEKSLFFSQAN AIAQDKTTGSVMLDNGVEIINDFRALKVEGASIPLKAFVDIESITNGKFY YNEIDSKAQIYLLFLREYKSFVILDESLYNSAYIQMFLLNQYDQDLFEQV TNDTRAKIYRLKR
SEQ ID NO: 2 lists the amino acids of the hexapeptide DQNATF{pNO2} (where F{pNO2} is paranitro-phenylalanine) representing the optimised oligosaccharide acceptor substrate for the OST (SEQ ID NO:1). This hexapaptide was crystallized together with the OST of C. lari PglB (SEQ ID NO:1) to give the atomic structure coordinates of Table 1 below, the statistics of which are provided in Table 2.
Table 1 is shown in FIG. 7. It lists the atomic structure coordinates of the crystallized oligosaccharyltransferase (OST, chain A) of Campylobacter lari (SEQ ID NO:1) complexed with peptide sequence DQNATF{pNO2} (SEQ ID NO: 2) (chain B), the optimal substrate for OST glycosylation and a bound divalent metal ion (chain C), useful for generating a three-dimensional model of the catalytic domain of the oligosaccharyltransferase (OST) of Campylobacter lari (SEQ ID NO:1). The table contains following information:
COLUMNSCONTENTS1-6Atom 7-11Atom serial number13-16Atom name17Alternate location indicator18-20Residue name22Chain identifier23-26Residue sequence number27Code for insertion of residues31-38Orthogonal coordinates for X in Angstroms39-46Orthogonal coordinates for Y in Angstroms47-54Orthogonal coordinates for Z in Angstroms55-60Occupancy61-66Temperature factor (Default = 0.0)73-76Segment identifier, left-justified77-78Element symbol, right-justified79-80Charge on the atom
Table 2 lists the X-ray data collection and refinement statistics for Table 1 in FIG. 7.
A. Data Collection Statistics
Data setNativeEMP1EMP2EMP3Beamline/detectorMD2 at SLSHighRes at SLSMD2 at SLSMD2 at SLSS06SA/PX1S06SA/PX1S06SA/PX1S06SA/PX1(Mar225)(Pilatus)(Mar225)(Mar225)SoftwareXDS/HKLXDSHKLHKLWavelength (Å)1.01.0 1.01.0Space groupP212121P212121P212121P212121Unit cell: a (Å)85.0685.586.187.8b (Å)116.1116.4117.0 119.4c (Å)175.04175.2174.8 169.9Resolution30-3.430-4.4530-3.830-4.2Crystal positions collected1234Completenes (%)99.3%(97.6%)99.6%(100%)92.3(68.3)96.9(84.2)Redundancy9.6(8.7)11.19.2(7.3)9.2(7.2)<I/□(I)>13.2(1.3)10.8(2.25)11.1(0.8)13.2(2.6)Rmgd-F (XDS) (%)13.9(132.8)15.2(86.3)Rmrgd (HKL) (%) 13.3a16(50.7)aNo Rmrgd factors indicated by HKL due to severe anisotropyB. Refinement Statistics (Native Data)
Resolution (Å)30-3.4No. of reflections working set (test set) 21834 (2000)Rwork/Rfree (%)23.8/27.1rmsd from idealitybond lengths (Å)0.011bond angles (°)1.475Average B factor (Å2)PglB129Peptide117Ramachandran analysis (Molprobity)Ramachandran favored82.6%Ramachandran outliers 1.5%